Spectrum 10K, the largest study of autism in the UK, was launched on August 24, 2021 amid a flurry of media attention—only to be paused less than three weeks later following an outcry from the autistic community (of which I am a member.) Concerns were raised around ethics and data management, but the primary issue is the potential use of such research for eugenic purposes.

I am peripherally associated with the biomedical autism research community, as a member of the AIMS-2-Trials Autism Representatives Steering Committee, and ECRAN, the associated early career researcher network. My involvement in such a controversial project stems from my desire to potentially influence the research agenda for the benefit of the autistic community: Many autism researchers have never known an autistic person, except as a research subject, and interacting as professional peers has been illuminating for many. (My perspective is outlined in a talk given at an ECRAN workshop in 2020 entitled Stigma and Unconscious Bias.)

As a medical doctor—a consultant anaesthetist—my knowledge of autism genetics and genomics research is basic. I mix up the terminology, and watch colleagues wince when I say genetics instead of genomics, although I know the difference. (I experience the same automatic wince when researchers use person-first language or misuse the term ‘neurodiverse.’)

My own research interests focus on healthcare access for autistic adults and issues facing autistic medical professionals, therefore I bring a range of perspectives as I consider my response to Spectrum 10K. I have sufficient knowledge of the current autism research agenda, the history of autism, and current advances in reproductive medicine to be seriously concerned about this project and similar research underway worldwide.

On the one hand the Spectrum 10K team states their values on their website, including a clear anti-eugenics stance:

“Spectrum 10K is anti-eugenics. To be precise, we do not work towards or advocate for termination of pregnancy in relation to a foetus who may later be diagnosed as autistic. We are aware of and abhor the awful history of eugenics, including how science has been used to justify many atrocities globally.”

On the other, this project is basic science dressed up as helping autistic people with co-occurring medical conditions. But ultimately it is a genomics project which will add to the evidence base around polygenic scores (PGS) for autism. These polygenic risk scores for medical conditions are already being used to select embryos during in vitro fertilization (IVF). There are already toddlers in this world who were selected as embryos based on polygenic scores for various medical conditions; the implications of this new direction in reproductive medicine were recently discussed in a leading medical journal. Such a test is not available for autism yet, but we have to ask if this is where the research is leading?

As for the AIMS-2-Trials, it is an international consortium of research teams and industry partners who are “exploring the biology of autism to tailor treatments and develop new medicines” and who also claim to oppose eugenics: “AIMS-2-TRIALS is not developing a genetic test and is ethically opposed to eugenics. It is also unlikely that a genetic test will ever be specific or accurate enough to predict whether a person will develop autism.”

While Spectrum 10K and AIMS-2-Trials are separate projects, there are researchers common to both. And both stake their anti-eugenics positioning on the expectation that a genetic test “specific or accurate enough to predict whether a person will develop autism” is unlikely. But are such claims disingenuous, given the current move towards using polygenic risk scores in reproductive medicine?

While such possibilities are denied, a preprint from a Spectrum 10K research team suggests otherwise. This study, not yet peer reviewed, suggests that polygenic scores (PGS) correlate with clinical profiles, and appear to show an apparent protective effect of autism PGS on co-occurring developmental disabilities. (This means that common genetic variants associated with autism are less likely to be associated with co-occurring intellectual disability, but ‘de novo’ or rare high impact genetic variants are more likely to be associated with intellectual disability.)

Recent and related discussions with leaders in the biomedical autism research field has led me to the astounding realisation: The possibility of paving pathways for eugenics may not be simply inadequately considered, but not considered at all. There is no attempt to hide the search for stratification biomarkers (biological features which can be used to group autistic people into subtypes), and it is unclear what the research community expects to be the outcome of such work.

I have no doubt that the individual researchers I have encountered are well-meaning and sincere in their desire to positively impact the world. However, there appears to be a total lack of awareness of current reproductive medicine practices’ bias against neurodivergent people, never mind the implications. For example, autistic people or even those with a family history of autism are currently prohibited from acting as sperm donors for IVF, with sperm banks subject to legal action for failing to prevent autistic children being born. Since 2013, it is accepted practice in Western Australia to use sex selection during IVF to reduce the likelihood of an autistic child (sex selection is not allowed in general, except where there is a history of autism).

Additionally, there is the ever-growing list of conditions that can be detected by preimplantation genetic diagnosis (PGD) and therefore prevented. Given the stigma associated with autism, and the focus on costs attributed to potentially life long care needs, the desire to add autism to that list seems to be driving much of the current research agenda.

Also, a recent, questionable paper by Blaxil et al claims that the increasing costs of autism poses a “serious threat to the economic future of the U.S,” and the authors propose a prevention strategy, citing a lack of evidence for reduced costs using an intervention strategy. Their “prevention” scenario assumes that strategies and opportunities are already “used by wealthy parents to lower their children’s risk of ASD” and suggest these “can be identified and made available rapidly to lower income children and ethnic minorities.” Criticisms of the paper are currently under consideration, but nonetheless, it was peer reviewed and published in Journal of Autism and Developmental Disorders.

So we know that society, if given a chance, will prevent us autistic people from existing. We know it is happening already, as detailed above. Is it possible, is it credible that biomedical autism researchers at the cutting edge of autism genetics research are genuinely unaware of developments in allied fields? Is this a result of working in silos, in which scientists are so isolated from those in other disciplines that they have no idea of how their works might intersect or overlap?

Spectrum 10K and similar projects will inevitably lead to polygenic scores being used to select for or against autistic people, particularly as data sharing alliances increase across the world. Therefore, as work on differentiating between groups of autistic people using polygenic scores and stratification biomarkers is ongoing, the obvious question arises: Is this an attempt to select for those who may be deemed useful, while preventing others with more complex needs from existing?

This is the essential question, the proverbial elephant in the room, as stratification is the subject of much current autism research. Yet this is not a straightforward quest. While on a superficial level the challenges associated with autism may appear to be split into separate “functioning” categories, the reality for autistic people is much more complex and fluid. Those who are familiar with the autistic world know this, yet the search for a genetic basis on which to segment our population continues.

Many autism researchers, however, are unfamiliar with the autistic world, and seem to assume that autism is to be avoided or prevented, even if it’s no longer considered acceptable to use the word “cure.” Or do researchers only avoid using that term in autistic company? What about in labs where autistic people are excluded, or forced to remain incognito? I hear valiant efforts to use appropriate terminology in spaces where we are known to be present, only to hear the same professionals revert to person-first language and functioning labels once back in the apparent security of their professional sphere. Are the considerations being shown to autistic concerns in Spectrum 10K and other autism research realms merely performative?

There is a cognitive dissonance that arises from the contradictions inherent in stating “we do not want a prenatal test for autism to be developed,” while simultaneously producing basic science which will inevitably lead to that outcome. As an anaesthetist, that is akin to saying I don’t want patients to be aware during surgery, while failing to administer anaesthesia. The inherent contradictions in this positioning are baffling. How is it credible that scientists working in this field are unaware of these developments in the wider field of reproductive medicine, as appears to be the case?

The alternative is that they are aware but are being disingenuous, and I don’t know if that is worse. I also have no clue whether individual scientists are being honest or not. It’s likely that there are some in either group. Is it possible that some within the biomedical research community are well-meaning but naïve, and also being manipulated by those with darker motives or commercially driven intent? Or are they so compartmentalized in their thinking that they see no inherent conflict in pursuing autism genetics research while espousing an anti-eugenics stance?

While ethical issues in genomics research are common across all specialties, medical conditions are generally acknowledged to be worth preventing. Curing cancer is in no way akin to eugenics, but autism is different. Autism is not an illness, and autistic people have the right to exist. Our community has already been the target of eugenics, such as Aktion T4 in Nazi Germany and the paediatric euthanasia programme at Am Spiegelgrund. Hans Asperger, once lauded as a hero, is now hated by many, following revelations he condemned to death those children considered to be unbearable burdens to their families. Yet how can those who exhort us to “learn from history and not repeat its terrifying mistakes” fail to appreciate that their current work may lead to similar outcomes? Do they genuinely differentiate between active killing of those autistic people deemed unworthy of life, and prenatal prevention?

Or is this the hidden agenda? Is there a tacit understanding among scientists that the actual goal is in fact to prevent the births of those deemed “undesirable,” while keeping those who might contribute to society? My heart breaks at this thought. Is there societal support for this potential outcome, driven by the cost of care needs?

Scientists being unaware of the potential for harm unleashed by their work being used in unanticipated ways is not new. Scientific regret is well documented, as exemplified by Oppenheimer, Einstein, and even Nobel, as regret was the catalyst for establishment of the Nobel Peace Prize. But the scientists involved in Spectrum 10K cannot claim ignorance. They have been told, loud and clear by the autistic community, from the earliest days of EU AIMS, Aims-2-Trials and since, culminating in the outcry over and boycott of Spectrum 10K.

Spectrum 10K is currently paused to “co-design and conduct a meaningful consultation.” The approach of setting a research agenda, then including autistic people in a tokenistic manner—as with AIMS-2-Trials A-Reps—is no longer sufficient. The autistic community, including autistic researchers and scientists, must be involved at the highest levels when decisions are made regarding the autism research agenda. We cannot be kept busy with token consultations while the real decisions are made elsewhere.

The gulf between the biomedical research and the autistic communities is far deeper than realised by the research community. An apology is needed for past harms to our community by biomedical research, but perhaps this needs to start with exploring what those harms have been. As we see in non-pharmacological autism interventions, the harms may not be obvious to non-autistic people. For example, alongside financial costs, mental health issues for autistic people are often suggested as a target of biomedical autism research, without any appreciation that the current intervention and research agenda may in fact be exacerbating mental ill-health.

As a community, autistic people have documented our research priorities, and causes of autism or “cures” do not feature prominently. We want research that will directly and positively affect our lives. We recognise that biomedical researchers generally do not have the skills to conduct the type of research we want, but the skills they do have can potentially be used for our benefit. Genomics offers enormous potential for improving human health. Knowledge in this area is accelerating at an astonishing rate, with debates raging around the ethics of genome editing, and regulation urgently required to avoid discrimination against vulnerable groups.

Autistic people are a vulnerable community, and as stakeholders in this debate, we deserve full and transparent engagement. We cannot allow the scientific process to continue in a direction that results in elimination of other autistic people, only for researchers to express regret when the inevitable outcomes occur. Scientists may claim that regulation is the role of society, not science. Yet who guides society on such matters, if not scientists with knowledge of the field?

These are difficult issues. If the obvious conclusion is valid; that PGS or other interventions may be used either to avoid autistic babies, or to avoid those deemed to have a higher likelihood of complex needs, then this must be openly acknowledged. If the biomedical research community is blindly forging ahead with lack of awareness or insight, this is an altogether different—but no less pressing—problem. In either case, the only way forward is absolute transparency, open debate, and a truly participatory approach to future research.